Author: Uri Blass
Date: 03:20:21 06/22/01
Go up one level in this thread
On June 21, 2001 at 23:45:21, Dann Corbit wrote: >On June 21, 2001 at 17:15:59, Sune Larsson wrote: > >> >> [D]4k2r/rp5p/pR4p1/4qp2/3pp1PQ/8/2P4P/R5K1 w k - 0 28 >> >> >> This position is from IM Berg-GM Hector, Sigeman tournament - Malmo. >> These two very tactical skilled Swedes had battled it out just like >> old masters did in the 19th century. Young Emanuel Berg sacked one >> pawn after another but, as it looked, in vain. Now, all good stories >> give some glory to the brave one and that is also what happened here. >> In the above position white, 4 pawns less, has the resource 28.Ra5!! - >> with the point 28.-Qxa5 29.Qf6! (not 29.Re6+ Kd7). At the present >> moment it looks like this 28.Ra5!! is good enough to draw the game >> for white. >> >> >> Test 1: Can your program find the move 28.Ra5!! - evals? > >It's a terrible [definitely losing] blunder, throwing away the rook for no >compensation: > >[D] 4k2r/rp5p/pR4p1/R3qp2/3pp1PQ/8/2P4P/6K1 b k - acd 15; acn 532177870; acs >1000; ce 1126; pv Qxa5 Rxg6 hxg6 Qxh8+ Kf7 Qxd4 Qe1+ Kg2 Qe2+ Kg1 Qxg4+ Kf1 Qf4+ >Kg2 Ra8 Qd5+ Kf8 c4 Qf3+ Kg1 Qe3+ Kg2 Qe2+ Kg1 Ra7; The big blunder is in the way that chess programs work. I am not surprised that they cannot see nothing. The reason is simply the fact that they do not use a good algorithm. Programs except gustav cannot see a mate in 70 that was posted here and again the reason is simply the fact that all programs except gustav do not use the right algorithm to find mate in 70. After 1...Qxa5 2.Qf6 black must play 2...Qe1+ to get big advantage by the evaluation of the programs. after 3.Kg2 black is again forced to play a check. The number of checks is very high and after enough checks programs do not extend enough and cannot see that black has to play checks. Uri
This page took 0 seconds to execute
Last modified: Thu, 15 Apr 21 08:11:13 -0700
Current Computer Chess Club Forums at Talkchess. This site by Sean Mintz.